Adsorptive iron removal from groundwater

(1)

3113

39 MMà

NN08200

40946 2001-12-17

% '•*.:•.£*$,,.••„*••<:'•.••, i , v : ; i;- v : ; - '*«S:'* : *'> -¥'•

Adsorptive Iron Removal

from Groundwater

(2)

Adsorptive Iron Removal from Groundwater

DISSERTATION

Submitted in fulfilment of the requirements of

the Academic Board of Wageningen University and the Academic Board of the

International Institute for Infrastructural, Hydraulic and Environmental

Engineering

for the Degree of DOCTOR

to be defended in public

on Wednesday, 19 December 2001 at 13:00 h in Delft, The Netherlands

by

SAROJ KUMAR SHARMA

born in Kathmandu, Nepal

(3)

Prof. dr. ir. J.C. Schippers, IHE Delft/Wageningen University, The Netherlands Members of the Awarding Committee:

Prof. dr. W. Rulkens, Wageningen University, The Netherlands Prof. dr. V.A. Snoeyink, University of Illinois, USA

Prof. dr. J.W. Geus, Utrecht University, The Netherlands Prof. dr. ing. R. Gimbel, University of Duisburg, Germany

Prof. ir. J.C. van Dijk, Delft University of Technology, The Netherlands

Prof. dr. ir. G.J.F.R. Alaerts, IHE Delft/Delft University of Technology, The Netherlands

All rights reserved No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without the prior written permission of the publishers.

(4)

THESES

belonging to the dissertation

ADSORPTIVE IRON REMOVAL FROM

GROUNDWATER

Saroj Kumar Sharma

19 December 2001

Wageningen University/ IHE Delft

The Netherlands

(6)

1. Adsorptive iron removal is economically attractive and

environmentally sound, compared to other available

techniques of iron removal.

2. The boundary between physicochemical and biological iron

removal is not well defined (Water Treatment Handbook:

Degremont 1991). Adsorptive iron removal has been very

often termed as 'biological' due to lack of explanations.

3. Despite human achievement in space technology, medical

science and genetic engineering, one of the basic problems of

human development is still providing "taps and toilets for all".

4. It is a paradox that people who develop or understand the

technology do not manage it, and those who manage the

technology do not understand it.

5. Science and technology dictate our culture, especially

languages and lifestyle. Nowadays people explore the world

with a mouse, carry palms in their pockets and produce babies

in test tubes. They outsource grocery shopping and kitchen

gardening but, like addicts spend their "limited free time"

indoors, glued to the 'box'.

(7)

they have to use the so-called 'best available technology'

marketed by the developed countries or the so-called

'appropriate technology' dictated by their economy.

7. There are two types of people: those who do the work and

those who take the credit. It is better to be in the first group,

because there is less competition.

8. The great end of life is not knowledge, but action. What men

need is as much knowledge as they can organise for action;

give them more and it may become injurious {Thomas Henry

Huxley). Some men are heavy with knowledge, yet still

(8)

Acknowledgements

Many people have contributed their time, energy, ideas, experience and encouragement to help me complete this study, for which I would like to take this opportunity to express my appreciation.

First and foremost, I extend my thanks to my supervisor Prof. Dr. ir. J. C. Schippers and my mentor Dr. ir. B. Petrusevski who were both instrumental in providing focus and channelling the study in a meaningful direction, and for providing valuable feedback at the critical junctures. Prof. Schippers was always inspiring and educating. I enjoyed working with him and learned a lot from his broad knowledge and vast experience. Dr. Petrusevski was very friendly, encouraging and always a source of help when needed.

A special gratitude to Dr. M. R. Greetham for his encouragement and guidance during the difficult initial period of this study. I also owe Ir. A. N. van Breemen and Dr. M. D. Kennedy my sincere thanks for their help and advice from time to time.

Gratefully acknowledged is the financial support of Kiwa N.V. Water Research and five Water Supply Companies in The Netherlands, namely: Water Supply North West Brabant (WNWB), Water Supply East Brabant (WOB), Water Supply Gelderland (WG), Water Supply Drenthe (WMD) and Water Supply Overijssel (WMO) to conduct this study.

Ir. M. Groenendijk and Ing. M. van der Haar of WNWB were very helpful in providing all necessary support during the pilot plant study and making available filter media from different plants from time to time. I would also like to thank Dr. G. IJpelaar, Dr. ir. B. Heijman, Ir. J. Kappelhof and Ir. G. Reijnen from Kiwa N.V. Water Research and Dr. E. Orlandini who were always encouraging and willing to answer questions and pass on what they had learned.

Prof. Dr. J. W. Geus and Ing. C. van Bennekom are acknowledged with thanks for their valuable suggestions as members of the Research Project Steering Committee. I am also thankful to all the members of the Contact Group Groundwater Filtration-Kiwa for their constructive comments and guidance during the course of this study. I am grateful to Ms. W. Sturrock for the proof-reading and English corrections of the draft of this thesis and to Ir. A. N. van Breemen for the Dutch translation of the abstract.

Thanks are also due to seven M.Sc. students involved in this study namely Mr. D.O. Owore, Mr. H. Mutikanga, Mr. R.A.B.S. Mendis, Mr. H. Karunatilake, Mr. C. Sebwato, Mr. D.K. Kironde and Mrs. N. Uyouko for their hardwork and valuable contribution.

(9)

My sincere thanks and appreciation to the IHE laboratory staff namely Mr. F. Kruis, Mr. F. Wiegman, Mr. C. Bik, and Mr. P. Heerings for their help and experimental support. I am also thankful to IHE staff members for their support during the study. Thanks to my colleagues Ingrida Bremere and Mustafa Mousa and all the Nepalese participants at IHE in the last six years for their direct and indirect help, encouragement and for providing good company, making my stay in Delft memorable.

Finally, I would like to thank my mother, my wife Sushmita and our son Sumeet for their continuous support, endless patience and understanding.

(10)

Abstract

Iron is commonly present in groundwater worldwide. The presence of iron in the drinking water supply is not harmful to human health, however it is undesirable. Bad taste, discoloration, staining, deposition in the distribution system leading to aftergrowth, and incidences of high turbidity are some of the aesthetic and operational problems associated with iron in water supplies. Iron removal from groundwater is, therefore, a major concern for water supply companies using groundwater.

Aeration followed by rapid sand filtration is the most commonly used method for the removal of iron from groundwater. Different mechanisms (physical, chemical, and biological) may contribute to the removal of iron in filters and the dominant mechanism depends on water quality and process conditions applied. Under the commonly applied treatment conditions in iron removal plants, the oxidation-floc formation mechanism (floe filtration) is commonly believed to be dominant. In this mechanism soluble iron(II) present in anoxic groundwater is oxidised to insoluble iron(III) and after precipitation, iron hydroxide floes are removed in the rapid sand filters. The second mechanism, adsorption-oxidation (adsorptive filtration) may play a role as well and has several potential advantages over the oxidation-floc formation mechanism, namely longer filter run, better filtrate quality, shorter filter ripening time, and less backwash water use and sludge production. In the adsorption-oxidation mechanism, the iron(II) is removed by adsorption onto the surface of the filter media. Subsequently, in the presence of oxygen, the adsorbed iron(II) is oxidised forming a new surface for adsorption. In conventional iron removal filters, the adsorption-oxidation mechanism is expected to be responsible for the removal of an important part of iron entering the filter bed in iron(II) form. Adsorptive filtration is most likely the dominant mechanism in dry filters and in sub-surface iron removal.

Water supply companies are continually seeking means to improve the process efficiency of iron removal from groundwater in order to minimise the deposition of iron in distribution networks, the backwash water use, and the volume of sludge produced. The WHO recommended guideline value of iron in drinking water is 0.3 mg/1 and the EC directive has set a parametric value of 0.2 mg/1. In the Netherlands, several water supply companies are aiming at an iron concentration <0.03 mg/1 in water supplies. Meeting these stringent requirements of iron in the water supply and backwash water treatment in developed countries, and reducing the operation and maintenance costs of distribution systems worldwide will require a more efficient removal and/or minimisation of the iron currently passing through the filters. Until now, it has been generally believed that, regardless of the water quality, the treatment approach was based on physical removal of the iron hydroxide floes. However, depending upon the water quality and process conditions applied, the application of adsorptive filtration may result in a higher process efficiency than floe filtration. A better understanding of the different mechanisms involved in the iron removal

(11)

terms of run time, filtrate quality, and overall treatment costs. Despite several advantages, the adsorption-oxidation mechanism has not been knowingly employed as the sole or dominant mechanism of iron removal in wet filters yet.

The goal of this research has been to examine the potential of adsorptive iron removal as an alternative to the conventional floe formation method and to investigate the factors governing the adsorptive iron removal process, particularly the mechanisms involved under anoxic conditions. This was accomplished by i) measuring the iron(II) adsorption capacities of several new filter media, iron oxide coated sand from iron removal plants, and iron hydroxides, ii) analysing the effect of water quality parameters on adsorption capacity, iii) investigating the effect of pH, iron concentration and filter media type on iron oxide coating development, iv) comparing the performance of pilot filters operating in floe filtration and adsorptive filtration modes, and v) modelling the adsorptive iron removal in filters to predict the iron breakthrough in filters with new and iron oxide coated sand under anoxic conditions. An experimental method was developed to measure the adsorption of iron(II) onto filter media and iron hydroxides. It was found that adsorption of iron(II) onto filter media can adequately be described with both the Freundlich and the Langmuir isotherms. The iron(II) adsorption capacities of the different filter media tested varied widely. Of the virgin materials tested, basalt showed the highest iron(II) adsorption capacity followed by anthracite, olivine, magnetite, sand, pumice, and limestone. Iron oxide coated sands from full-scale iron removal plants demonstrated a much higher capacity for iron(II) adsorption compared to new (virgin) sand. In the pH range examined (6-7.5), the iron(II) adsorption capacity of both new and iron oxide coated sand increased with the increase of pH. Among the iron hydroxides tested, lepidocrocite had the highest iron(II) adsorption capacity, followed by amorphous iron hydroxide, ferrihydrite, and goethite. An estimation based on the experimental results indicated that the contribution of iron(II) adsorption onto iron hydroxide floes on the overall process of adsorptive iron removal, as well as floe filtration iron removal, is probably negligible in iron removal plants.

The high iron(II) adsorption capacity of iron oxide coated sand from iron removal plants can, in principle, be utilised to improve iron removal in filters by switching the governing mode of operation from floe filtration to adsorptive filtration. This can be achieved by bypassing the aeration step and/or reducing the pre-oxidation time to ensure that the majority of the iron enters the filter bed in iron(II) form. In practice, primarily adsorptive iron removal in filters can be realised in two operational modes, namely a) intermittent regeneration mode by operating the filter under anoxic conditions and regenerating the adsorption sites by e.g. backwashing with oxygen-rich water or with a chemical oxidant e.g. KMn04, and b)

continuous regeneration mode by operating the filters under aerobic conditions with limited oxygen concentration in the feed water and/or by limiting the pre-oxidation time.

Within the concentration range examined, NIL»*, CI', alkalinity, and background ionic strength had no significant effect on iron(II) adsorption onto either new silica sand or iron

(12)

XI

óxide coated sand. An increase in concentrations of Mn2+, Ca2+, PO43", and silica decreased

the iron(II) adsorption onto new sand, whereas an increase in iron(II) adsorption was observed when SO42" concentration was increased. The effect of Mn2+, Ca2+, SO42" and PO43"

on iron(II) adsorption onto iron oxide coated sand, however, was minimal. In general, the effect of different parameters on iron(II) adsorption was more pronounced on new sand than on iron oxide coated sand. This indicates that once the iron oxide coating is developed on the surface of the filter media, iron(II) adsorption is not hindered significantly by the presence of other inorganic ions within the concentration range common for groundwater. Preliminary experiments with commercial humic acid showed a negative effect of organic matter on iron(II) adsorption capacity. The effect of organic matter present in groundwater on iron(II) adsorption onto filter media needs further detailed investigation.

Analysis of the physical and surface chemical characteristics of coated sand from twelve groundwater treatment plants in the Netherlands showed that compared to new sand, coated sand had a very high porosity and a very large specific surface area. In general, the iron content of the coating and iron(II) adsorption capacity increased with time in use. However, the average annual increase of the iron content and the adsorption capacity varied for the coated sand from different plants, probably due to the difference in water quality, process conditions applied, and time in use. The grain size of the filter sand increased and the density decreased with the development of iron oxide coating. The decrease in density of coated sand with the iron oxide coating development was a function of the increase in the effective grain size. The measured high adsorption capacities of coated sand from wet filters and dry filters of full-scale groundwater treatment plants indicate that, in wet filters, adsorptive iron removal also plays a role. In dry filters, this mechanism should be dominant due to a very short pre-oxidation time.

The development of an iron oxide coating on the filter media is an important factor in effective adsorptive iron removal from groundwater. The rate of development of the coating and its characteristics are influenced by raw water quality, process conditions applied, and characteristics of the filter media. It was found that preconditioning of new filter media (e.g. at high feed water pH and/or high iron concentration) results in rapid development of an effective iron oxide coating that can reduce the initial filter ripening time. Additionally, the use of virgin media with a high iron(II) adsorption capacity, like basalt, can also reduce the time required to develop a coating with an adequate adsorption capacity.

The process of adsorptive iron removal in filters under anoxic conditions was modelled using adsorption isotherm parameters, mass balance, and mass transfer equations. Experimental results were compared with the predictions of three fixed bed adsorption models, namely i) Constant Pattern Model (CPM), ii) Linear Driving Force Model (LDFM), and iii) Plug Flow Homogeneous Surface Diffusion Model (PFHSDM). The CPM, which considers external mass transfer only, was not sufficient to predict iron breakthrough in filter columns with new and iron oxide coated sand. The LDFM and the PFHSDM predictions of iron breakthrough were more accurate in the case of new sand. In the case of iron oxide coated sand, the predictions were not satisfactory. The difference in model predictions and experimental

(13)

results in the case of iron oxide coated sand was probably due to the effect of an initial pH drop in the pores of the filter media with iron(II) adsorption, and a consequent decrease in iron(II) adsorption capacity.

Adsorptive filtration can potentially be employed as the primary method of iron removal from anoxic groundwater without manganese and ammonium. This process could also be very attractive in situations where two filtration steps are applied due to high concentrations of iron, manganese, and ammonium in raw water. The first filter can be optimised as an adsorptive iron removal filter, while the second filter can be employed for manganese and ammonium removal.

Adsorptive iron removal is potentially an attractive alternative to conventional floe filtration iron removal. Application of this process has prospects of improving the filtrate quality, extending the filter run time, and easing or reducing the treatment of filter backwash water and sludge, thus resulting in a higher treatment process efficiency.

Key words: groundwater, iron removal, filtration, removal mechanisms, floe formation,

adsorption, process efficiency, adsorption capacity, iron oxide coated sand, water quality, coating development, modelling, mass transfer.

(14)

Chapter 1 INTRODUCTION

1.1 Groundwater use and treatment 3 1.1.1 Groundwater for water supply 3

1.1.2 Iron in groundwater 4 1.1.3 Groundwater use and iron removal practice in The Netherlands 6

1.1.4 Groundwater use and iron removal practice in Nepal 7

1.2 Iron removal methods 8 1.3 Chemistry of iron removal 12

1.3.1 Iron oxygenation kinetics 12 1.3.2 Factors affecting oxidation of iron 15

1.3.3 Hydrolysis of iron(III) 16 1.3.4 Chemical oxidation 17 1.4 Iron removal mechanisms in filters 18

1.4.1 Oxidation-floc formation 18 1.4.2 Adsorption-oxidation 19 1.4.3 Biological iron removal 20 1.5 The need for the research 23 1.6 Research objectives 24 1.7 Outline of the thesis 24

(15)

1.1 GROUNDWATER USE AND TREATMENT 1.1.1 Groundwater for water supply

Groundwater has been used as a source of drinking water since time immemorial. In Egypt wells were already being used in 3000 BC (Katko 1997). Excavations at Mohanjodaro have revealed brick-lined dug wells existing as early as 3000 BC during the Indus Valley Civilisation (Raghunath 1987). Wells and their importance can also be traced in the Book of Genesis. The Bible recounts numerous incidents which illustrate the importance of groundwater supplies to the tribes of Israel (Rail 1989).

Groundwater is the major source of drinking water in many countries across the world. Table 1.1 summarises the groundwater use as drinking water in different regions. Groundwater is extensively used as an important source of public water supply in Europe ranging from nearly 100% in Denmark, 72% in Germany, and 56% in France to 27% in the United Kingdom (EEA 1999). In the United States, groundwater is the primary source of potable water for over 96% of the rural population (Biswas 1997). In some Asian countries the share of groundwater in drinking water supplies was as follows: India 80% (rural), Maldives 80%, Philippines 60%, Thailand 50%), and Nepal 60% (Das Gupta 1991). These inventories of groundwater use in water supply reveal its worldwide importance.

Table 1.1 Groundwater as a share of drinking water by region

Region Asia Pacific Europe Latin America United States Australia World Shan From ; of drinking water i groundwater (%) 32 75 29 51 15 People served (millions) 1000 to 1200 200 to 500 150 135 3 1500 to 2000 Source: Sampat (2000)

Groundwater is generally a preferred source for water supplies because of its convenient availability close to where water is required, its constant and good natural quality (which is frequently adequate for potable water supplies with minimal treatment), and relatively low capital cost of water supply system development. Against these common advantages, it should be noted that groundwater is vulnerable to contamination by various anthropogenic activities (agricultural, domestic, and industrial). Contrary to the popular impression that the waters from springs and wells are "pure", patterns of pervasive pollution of groundwaters are being uncovered.

(16)

Adsorptive Iron Removal from Groundwater

Groundwater is naturally of very good microbiological quality and its chemical quality depends on hydrogeological conditions. Naturally occurring groundwater quality problems are typically associated with high hardness, high salinity and elevated concentrations of iron, manganese, ammonium, fluoride, and occasionally methane, hydrogen sulphide, nitrate, and arsenic (Nash and McGall 1994). Hence, generally some form of treatment will be required for potable water supplies. A high concentration of iron and manganese is by far the most common water quality problem associated with groundwater.

1.1.2 Iron in groundwater

Iron being the fourth most abundant element and second most abundant metal in the earth's crust (Silver 1993; WHO 1996), is a common constituent of groundwater. The presence of iron in groundwater is generally attributed to the dissolution of iron bearing rocks and minerals, chiefly oxides (hematite, magnetite, limonite), sulphides (pyrite), carbonates (siderite) and silicates (pyroxene, amphiboles, biotites and olivines) under anaerobic conditions in the presence of reducing agents like organic matter and hydrogen sulphide (O'Connor 1971; Hem 1989). Iron usually exists in two oxidation states, reduced soluble divalent ferrous (Fe2+or iron(II)) and

oxidised trivalent ferric (Fe3+or iron(III)). Iron may be present in groundwater in the following

five forms: i) dissolved as iron(II), ii) inorganic complexes, iii) organic complexes, iv) colloidal, and v) suspended. The state of the iron in water depends above all on the pH and the redox potential (Eh) (Fig. 1.1).

Most natural waters have pH values ranging from 5.0 to 8.5, and pE values ranging from -7 to +12. Thus, iron(II) would be the predominant species in the absence of an electron acceptor such as oxygen (Hem 1989; Faust and Aly 1998). The concentration of iron in natural waters is frequently limited by the solubility of its carbonate. Waters of high alkalinity often, therefore, have a lower iron content than water of low alkalinity (O'Connor 1971; ASCE and AWWA

1990). Iron concentration in groundwater normally ranges from a few hundredths to about 50 mg/1 with the majority containing <5 mg/1 (Hem 1989; Davis 1997).

(17)

pH +20 +16 •12 PE » 10 12

i i r

J« WATER OXIDIZED _ WATER REDUCED

miiiimii/iiHi'HuiiiniiiKHW iimiiiiiiuuwiiiiiiinnutiuiiiiwiiiiHi

Fig. 1.1 pH-pE diagram for the iron system (Faust and Aly 1998)

The iron problem

There is no health consequence of iron in drinking water. Iron is an essential element in human nutrition. Estimates for the minimum daily requirement for iron depend on age, sex, physiological status and iron bioavailability and range from 10 to 50 mg/day. An intake of 0.4-1 mg iron/kg of body weight per day is unlikely to cause adverse effects in healthy persons. Allocation of 10% of this provisional maximum tolerable daily intake (PMTDI) to drinking water gives a value of about 2 mg/1, which does not present a hazard to health (WHO 1996).

Iron is normally present in groundwater worldwide. Iron in water supplies, however, is undesirable, as it is a nuisance for domestic and industrial users and water suppliers causing various aesthetic and operational problems as listed below.

(18)

Adsorptive Iron Removal from Groundwater

1. Iron produces ugly and insoluble rusty oxide-red, yellow or brown stains and streaks on laundry and plumbing fixtures (O'Connor 1971; Kothari 1988). In extreme cases, iron interferes with the culinary use turning tea black and darkening the boiled vegetables (Hauer

1950).

2. Iron imparts colour and a typical bitter, astringent taste to the water. The taste threshold of iron in water is 0.04-0.1 mg/l (JMM 1985; WHO 1996). Turbidity and colour may develop in piped systems at iron levels above 0.05-0.1 mg/1 (WHO 1996). Though harmless, these organoleptic characteristics give the impression that the water is somehow contaminated. Most importantly in the developing countries, the colour and bitter taste caused by iron can result in well water being rejected. People then often return to the polluted surface water and so incidents of cholera and typhoid continue (Ahmed and Smith 1987, 1988; Chibi 1991). 3. The presence of iron is disastrous in some industrial wet processing operations. Water to be

used in the textile, dyeing, beverage and white paper industries should contain less than 0.05 mg/1 of iron or manganese (Cox 1964). Additionally, the oxidation of iron-rich water applied to cultivated fields can lead to low-pH ferric hydroxide-rich soils that may severely damage agricultural productivity (Chapelle 1993).

4. Iron passing into the distribution system may promote the growth of micro-organisms. Slime thicknesses of several centimetres have been observed in distribution pipes. These accumulations, consisting of hydrous iron and manganese oxides and bacteria, increase the friction loss and power consumption, require higher chlorine dosage, deplete dissolved oxygen, reduce the carrying capacity and may eventually clog the distribution pipes. Sloughing or resuspension of this material by high flow causes high turbidities. (O'Connor

1971; Culp 1986; Salvato 1992; Vigneswaran and Visvanathan 1995). Therefore, for the water supply companies the main concerns are minimising the costs of operation and maintenance and reducing the "red water" incidents.

Iron removal from groundwater is, therefore, a major concern for most water supply companies using groundwater as their source. To prevent the difficulties mentioned above, various regulatory agencies have put forward standards or guidelines to control iron concentrations in water supplies. An AWWA task group suggested limits of 0.05 mg/1 for iron and 0.01 mg/1 for manganese for an "ideal quality water" for public use (Bean 1962). Based on taste and nuisance considerations, the World Health Organisation (WHO) recommends that the iron concentration in drinking water should be less than 0.3 mg/1 (WHO 1996). The EC directive recommends that the iron in water supplies should be < 0.2 mg/1 (EC 1998). In the Netherlands, the guideline level for iron in drinking water is < 0.05 mg/1 (VEWIN 1993) and several water supply companies are aiming at a level of <0.03 mg/1 in order to minimise the distribution system maintenance costs.

1.1.3 Groundwater use and iron removal practice in the Netherlands

The kingdom of the Netherlands is situated along the North Sea in north-west Europe covering a land area of approximately 34,000 square kilometres. The total territory, including inland lakes, estuaries and territorial sea, amounts to 41,160 square kilometers. The topography is relatively flat with the elevation ranging from -6.7 m to 322 m above mean sea level.

(19)

The population of the Netherlands is about 15.9 million (2000), with an average population density of 460 per square kilometer, of which 99% have access to public water supply services (CBS 2001).

In the Netherlands, there are at present about 20 water supply companies with a total of 250 treatment facilities and an annual production of 1300 x 106 m3. A total of 232 groundwater

treatment plants, of which five abstract a mixture of groundwater and riverbank infiltrate, produce 805 x 106 m3/year (van der Kooij et al. 1999). Groundwater contributes to about 62%

of the water produced by water companies in the Netherlands (CBS 2001). The removal of iron, manganese, ammonium and methane is the most important step in groundwater treatment. Groundwater extracted in the Netherlands contains 0-30 mg/1 Fe (mean 4.8 mg/1), 0-2 mg/l Mn (mean 0.2 mg/1) and 0-35 mg/1 NH41" (mean 0.6 mg/1). The removal of these constituents is

carried out by a combination of aeration and filtration. Several aeration systems, namely: cascades, spray aeration, tower aeration (co-current and counter current), venturi aeration, and plate aeration are used for iron oxidation. Some plants also use potassium permanganate and ozone as oxidant, and aluminium or iron salts as coagulant to facilitate iron removal. The single or dual media rapid filters used are of open gravity, pressurised, and dry type (van Wijk et al.

1987; Kruithof and Koppers 1989). Many groundwater treatment plants employ two filtration steps for iron, manganese and ammonium removal. Iron is removed in the first filter and then manganese and ammonium are removed in the second filter. In the Netherlands, treatment of the filter backwash water is gradually becoming a common practice and a beneficial application of the sludge after thickening is now being introduced.

1.1.4 Groundwater use and iron removal practice in Nepal

Nepal is a landlocked, mountainous developing country in South Asia covering an area of 147,181 square kilometres. It is characterised by diverse physiographic zones, contrasting climates and attitudinal variations ranging from 75 to 8848 m. High mountains and rolling hills account for 83% of the total land area and the remaining 17% are occupied by the plains of Terai (CBS 1998). The estimated population in 2000 is 22.9 million and the growth rate is 2.4% per annum (MOPE 2000). Despite the high priority awarded to water supply and sanitation in Nepal and the rising public investments over the last decade, major shortfalls in the level and quality of service coverage still remain. The estimated national coverage of water supply services by the end of July 2000 is 61% (DWSS 2000).

A survey conducted by ESCAP in 1990 revealed that in Nepal 60% of drinking water, 80% of the municipal water supply and 20% of agricultural water supply comes from underground aquifers (Das Gupta 1991). Groundwater is extensively used for water supply in the Terai plains of Nepal and in the Kathmandu valley. In the Terai region, shallow hand pumps and deep tube wells are used extensively for the provision of drinking water. Iron in groundwater is a major quality concern in many rural areas where hand pumps and deep wells are used. People often complain of the bitter taste and the colour of the water. Some people still go back to the traditional surface water sources to collect drinking water and use water from wells for purposes

(20)

8 Adsorptive Iron Removal from Groundwater

other than drinking. Hand pumps and tube wells frequently fail to provide the required service due to clogging by iron deposits.

Attempts have been made to solve the problem in some areas by installing low-cost iron removal units (with a perforated pipe for aeration and twin chambers for sedimentation and filtration) promoted by UNICEF (UNICEF 1986). However, they are very few in number and the majority of them do not function properly or are not in use because of various operational problems. Some households practice domestic iron removal using three pitcher-sand filtration. Very often, hand pumps are abandoned or rejected shortly after installation by the government or donor agencies for other water sources with inferior water quality because of the colour and taste associated with groundwater from hand pumps. Thus, in the absence of proper iron removal methods, many water supply systems installed have failed to give the intended health benefits of safe drinking water.

Groundwater provides for about 40% of the water needs of Kathmandu valley towns (population about 1.1 million) during the dry season with the annual average contribution of about 25% (JICA 1990; ADB 1997; BTW 1998). Integrated surface water-groundwater treatment using the extensive treatment processes of bio-filtration, flocculation/sedimentation, and rapid sand filtration has been employed in Mahakalchaur and Bansbari water treatment plants in Kathmandu. In some Terai towns, namely Rajbiraj, Lahan, Damak, Inarwa, Kakarbhitta, Birtamod, Chandragadhi and Bhadrapur, some iron removal plants have been constructed which employ the process of aeration/chlorination followed by rapid filtration. Problems associated with iron are, however, often reported due to operational inconveniences. In many cases, the raw groundwater is often pumped directly into the distribution network. The lack of knowledge of the mechanisms of iron removal and the relatively high cost and complexity of providing the necessary treatment has led to either inadequate or no treatment.

1.2 IRON REMOVAL METHODS

Treating groundwater to remove iron from municipal, agricultural and domestic wells is a multi-million dollar a year business throughout the world (Chapelle 1993). The first iron removal plant was constructed at Charlottenburg, Germany in 1874. The earliest plants employed aeration and filtration, sometimes supplemented by the addition of lime, to treat groundwaters (O'Connor

1971). The same method of treatment predominates today.

Removal of iron from groundwater can be accomplished in several ways. The type of treatment largely depends on the quality of the raw water, financial resources available and the philosophy of the water company. The following methods are used to control iron in the water supply:

1. Oxidation-precipitation-filtration

(21)

the addition of chemicals for pH adjustment.

(b) Chemical oxidation using chlorine, chlorine dioxide, ozone, potassium permanganate or hydrogen peroxide (with or without pre-aeration).

(c) Biological oxidation 2. Ion exchange or zeolite softening

3. Stabilisation or sequestering using silicates or polyphosphates 4. Lime softening or limestone bed filtration

5. Manganese greensand process

6. In situ oxidation (subsurface iron removal or VYREDOX process) 7. Membrane processes

8. Calcined magnesite - diatomaceous earth filtration (O'Connor and Benson 1970) 9. Sirofloc (activated magnetite) process (Gregory et al. 1988; Home et al. 1992) 10. Catalytic or adsorptive filtration using patented filter media impregnated with

various oxides of iron and/or manganese like BIRM, PYROLOX, Anthrasand, Pyrolusite, Aqua Mandix, Catalytic Carbon, etc. (Sommerfeld 1999)

The suitability, advantages and limitations of some of the most commonly used methods are summarised in Table 1.2

Table 1.2 Iron Removal Methods

Removal Method 1. Oxidation, prec (a) Oxidation by aeration (b) Oxidation with chlorine Application

ipitation and filtration • Fe <5 mg/1 and little or

no organic matter or other reducing agents • As pre-oxidation step to

save chemical costs when Fe >5 mg/1

• Beneficial to remove Fe and Mn in one step • Optimum pH 6.8-8.4 (Kothari 1988) Advantage • No chemicals required • Simple in operation • Partly removes C02, H2S and CH4 present • More rapid oxidation than aeration especially under conditions of organic matter interference • Less expensive and more effective than KMn04 • Can also be used

for disinfection

Limitation

• Ineffective in cases of low pH, and high Fe and Mn

concentrations or when Fe is organically complexed • Initial cost is high • THM formation • Chloro-derivatives

can cause taste and odour problems • Require safe handling

and storage of chlorine and chlorine compounds

(22)

10 Adsorptive Iron Removal from Groundwater Removal Method (c) Oxidation with chlorine dioxide (d) Oxidation with potassium permanganate (e) Oxidation with Ozone (f) Oxidation with hydrogen peroxide (g) Biological iron removal Application

• Effective when iron is organically complexed or ammonium concentration is high (Twort et al. 2000) • Reacts more rapidly

with Mn than chlorine • Fe <5 mg/1 (Wong

1984)

• More efficient at pH >7.5

• Effective even when the iron is organically complexed (Cromley and O'Connor 1976; Paillarde/ al. 1991)

• Very effective when iron is organically complexed • Recommended for groundwater with acidic or neutral pH (Mouchet 1992; Bourginee/a/. 1994) Advantage • No THM formation • Less equipment and capital investment compared to chlorine • Efficient; rapid and complete reaction • Reacts with H2S, cyanides, phenols, and other taste and odour-producing compounds • Powerful and effective oxidant • Multi-purpose applications of ozone e.g. disinfection, color removal, taste and odour control • No THM

formation • Faster oxidation • Forms dense,

easily settled solid • Cheaper than

ozone

• Leaves no residue • Higher filtration

rate

• Longer filter run • Reduced capital and O&M costs

Limitation

• Costlier than chlorine • Possible health effects

of by-products (Twort et al. 2000)

• Not used for iron removal only (Culp

1986)

• Difficult to control • Overdose {= 0.05

mg/1) may produce a pink colour • Ineffective for high

iron concentrations • More expensive than

chlorine and ozone

• High initial capital and operating costs • May oxidise Mn2+ to Mn04" resulting in a pink colour • Formation of unwanted by-products e.g. Br03" • Formation of AOC • Sensitive to process conditions (pH and temperature dependent) • Ineffective in presence of NH4+ and H2S

(23)

Removal Method 2. Ion exchange (Zeolite softening) 3. Stabilisation or Sequestering process (with polyphosphates and silicates) 4. Lime softening/ Limestone bed filtration 5. Manganese greensand process Application

• Suitable for individual water supply scheme with<5 mg/lofFe (Gass 1977)

• Removes dissolved Fe and Mn together with hardness

• Used as a polishing step in some plants/ household use after municipal treatment

• Fe should be <1 mg/1 (Salvato 1992) • For distribution system

corrosion and deposition inhibition • Practical in controlled

use only

• Pre-aerated water with pH >9.5 and sufficient alkalinity (>20 mg/1 as CaC03). • Removes Fe and Mn by combination of sorption and oxidation • MaxFe + Mn<10mg/l • H2S <2-5 mg/1 • Optimum pH 6.2-8.0 Advantage • Softening occurs with exchange of Ca2+ and Mg2+ • Complexes iron and holds it in solution and the consumers do not notice its presence • No sludge generation • Beneficial when a large amount of softened water is required • H2S can be removed together with Fe and Mn Limitation • Possibility of resin/zeolite fouling or loss of exchange capacity in presence of C^due to iron precipitation • High capital cost • Requires skilled

personnel • Ineffective for

colloidal or complexed iron • More expensive than

Cl2 and KMn04 • Phosphate introduced may stimulate biological growth; May require chlorination to prevent bacterial growth

• Cold water only (complex break down when heated) • Polyphosphate

complexes may degrade after 48-72 hours (Lorenz et al.

1988)

• Not cost effective unless lime treatment is also required for hardness reduction (Culp 1986) • Increased sludge

problems • Higher head loss,

shorter run time • High O&M costs

(KMn04 regeneration) • Not suitable for larger

water treatment plants (JMM 1985)

(24)

12 Adsorptive Iron Removal from Groundwater Removal Method 6. In-situ oxidation (sub-surface removal/ Vyredox method) 7. Membrane processes Application • By infiltrating oxygen-rich water into the ground through a well

• NF/RO -toremove dissolved iron • MF/UF - to remove iron

floes Advantage • No chemicals • Abstraction-infiltration ratio of groundwater is high • Can be combined with the removal of other constituents e.g. hardness and THM precursors Limitation • Potential for contamination of aquifer • Clogging of the aquifer, corrosion of well screen • Excessive bacterial

growth may occur around the well. • High capital and

O&M costs • MF/UF membranes

require frequent cleaning

Among the different techniques mentioned above, aeration or chemical oxidation followed by rapid sand filtration is most widely used (O'Connor 1971; Wong 1984; JMM 1985; Culp 1986; Salvato 1992; Twort et al. 1994; Sommerfeld 1999). Aeration - rapid sand filtration is the preferred method in the Netherlands and in developing countries. Compared to other methods, this method is more economical, less complicated and generally avoids the use of chemicals, which is not usually welcome in the water industry.

1.3 CHEMISTRY OF IRON REMOVAL 1.3.1 Iron oxygenation kinetics

Iron oxidation and its removal is based on the transformation of the soluble form of iron (Fe ) to an insoluble form (Fe3+). In simplified notation,

4Fe2+ + 02+ 2H20 -> 4Fe3++ 40H

4Fe3+ + 40K + 8H20 -> 4Fe(OH)3+ 8IT

2+

4Fe + O2+10H2O-> 4Fe(OH)3+ 8lf

(1.1) (1.2) (1.3) Equation 1.1 shows that about 0.14 mg of oxygen is required for the oxidation of 1 mg of

iron(II). Therefore, the oxygen concentration in aerated water is theoretically sufficient for the complete oxidation of iron(II) normally present in natural groundwaters.

(25)

It is important to note that iron hydroxides formed after oxidation of iron(II) and subsequent hydrolysis of iron(III) is a complex of different iron hydroxide species and the representation of Fe(OH)3 is merely a simplification of reality. The hydrolysis of iron(III) is discussed in detail in

section 1.3.3.

Iron oxygenation kinetics has been extensively studied (Stumm and Lee 1961; Lerk 1965; Jobin and Ghosh 1972; Olson and Twardowski 1975; Sung and Morgan 1980; Robinson et al. 1981; Davison and Seed 1983). Stumm and Lee (1961) found that the rate of oxygenation of ferrous iron in bicarbonate solutions is of the first order with respect to both the concentrations of iron(II) and dissolved oxygen and of the second order with respect to the OH" ion.

-d[Fe(II)]

dt k0 p02 [Fe(II)] [OH' J2 (1.4)

where d[Fe(II)]/dt = rate of iron(II) oxidation (mol l"1 min"1)

ko = reaction rate constant = 8.0 (± 2.5) x 101312 mol"2 atm"1 min"1 at 20.5°C

pÛ2 = partial pressure of oxygen (atm) = 0.21 [02]/[02-sat]

[O2], [02-sat] = actual and the saturated concentration of oxygen in water (g/m3).

[Fe(II)] = concentration of ferrous iron (mol/1). [OH"] = concentration ofhydroxyl ions (mol/1)

Equation (1.4) shows that the oxygenation rate is very strongly pH-dependent, increasing 100-fold for each unit increase in pH. Therefore, the rate of oxidation of iron(II) is slow at low pH. Their studies also showed that oxidation of ferrous iron should be expected to occur rapidly in well-oxygenated waters at pH values exceeding 7.2.

20 30 Time, min

Fig. 1.2 Oxygenation rate of ferrous iron is proportional to Fe(II) and is strongly influenced by pH (Stumm and Lee 1961)

(26)

Recent studies have shown that in the presence of iron(III), the oxidation of iron(II) takes place via two parallel paths. One of these is the homogenous reaction occurring in the solution and the other is the heterogeneous reaction occurring on the surface of iron hydroxide precipitates. (Tamura et al. 1976; Tufeckci and Sarikaya 1996). At constant pH and 02 concentration, the rate

equation is given by

-d[Fe(II)] = (k + k<Fe[ni])[Fe(II)] (1.5)

dt

where k = rate constant for the homogeneous reaction = ko [O2] [OH"] k' = rate constant for the heterogeneous reaction = kso [02]K/[H ]

ko and kso are the real rate constants for the reactions and K is the equilibrium constant

for the adsorption of iron(II) on iron(III) hydroxide. The numerical values of the constants are ko = 2.3

(Tamura et al. 1976).

constants are ko = 2.3 x 101413 mol"3 s"1, kso = 73 1 mol"1 s"1 and K = 10'9'6 mol l"1 mg"1

The reaction is, therefore, autocatalytic as the oxidation of iron(II) is facilitated by the reaction-product iron hydroxides. The effect becomes noticeable at iron(III) concentrations exceeding 5-10 mg/1 (Tamura et al. 1976; Sarikaya 1980) and the oxidation rate reaches a maximum at iron(III) concentration of about 600 mg/1 (Tufekci and Sarikaya 1996).

At near neutral pH, most of the iron(III) is in the form of hydroxide precipitate with a positive surface charge. Consequently OH" is attracted into the diffuse layer; therefore, the pH of the diffuse layer of iron(III) floes is higher than that of the bulk solution. Thus iron(II) adsorbed on the surface of iron(III) precipitate is oxidised at much higher rates since it is known that the oxidation rate is proportional to the squares of [OH"] concentration (Tamura et al. 1976; Sarikaya 1980; Tufekci and Sarikaya 1996). This could offer an explanation for higher iron removal efficiency in the presence of iron(III) precipitates or iron oxide coatings on the media.

Barry et al. (1994) presented the general rate expression for iron oxidation kinetics considering homogeneous oxygenation, abiotic heterogeneous catalysis, biotic oxidation process and other mechanisms.

-d[Fe(II)l

dt = {k0[Fe2+] + k,[Fe(OHr] + k2[Fe(OH)2]}pO2

+ k'3 A[Fe(U][OH-]2p02 + k4 [Bacteria] [Fe(II)][OH' ]2 p02 +Rolher (L 6)

where

ko-k2 =

k'3 =

A =

first order homogeneous rate constants in water, adjusted for the presence of ligands and catalysts active in the homogeneous oxidation process (s"1 atm"1)

overall rate constant for heterogeneous, abiotic processes (l2 mol"2 m"2 s"1 atm"1)

(27)

k.4 = rate constant for some bacteria, the term repeated for various strains and corrected for

specific conditions (units are l3 mol"2 s"1 atm"1 cells"1 for bacteria measured in cells/1

or l3 mol"2 s"1 atm"1 g"1 for bacteria measured as grams volatile solids/1) [Bacteria] = concentration of bacteria (cells/1 or grams volatile solids/1)

Rother = the oxidation rate attributable to processes not considered explicitly, such as the

presence of reactive species like hydrogen peroxide (mol l"1 s"1)

This suggests that homogeneous oxygenation of iron may often be overshadowed by heterogeneous, biotic and photochemical mechanisms.

1.3.2 Factors affecting oxidation of iron

Besides pH, other water quality parameters like alkalinity (bicarbonate concentration), temperature, organic matter and some elements/ions have also been reported as having a significant effect on the rate of oxidation of iron(II).

Alkalinity

Alkalinity is important in iron removal as it provides the buffer capacity to avoid excess pH drop on iron oxidation and influences the characteristics of the precipitate formed. Iron oxidation and removal is poor at low alkalinity due to slow oxidation and poor floe formation (Robinson and Breland 1968; Hult 1973). Ghosh et al. (1966) reported that in groundwaters with high alkalinity (>250 mg/1 as CaCC^) the precipitates formed after aeration are primarily carbonates rather than hydroxides, and a large part of the iron precipitated is in the ferrous rather than the ferric form. Cleasby (1975) indicated that the more rapidly the iron is oxidised, that is, through the use of strong oxidants such as permanganate, chlorine or ozone, the more likely it is that the end product will be hydroxide. However, when the oxidation proceeds more slowly with aeration, then most likely the end product will be carbonate in water of high alkalinity.

Jobin and Ghosh (1972) found that buffer intensity (ß in eq/pH) of water influences the rate of iron oxidation at values higher than 4.0 x 10"3 eq/pH and suggested the following rate equation:

~d[FfI)] = k p02 [Fe(II)] [OH-]2[$]" (1.7)

dt

where ß =2.3 {[H+] + [OH"] + CT [<x,(a0 + a2) + 4a2a0]}

CT = [H2C03] + [HCO3-] +[C032"]

[H2C03] [HCO3-] . [C032-]

a0= * , «1 =L r and a2

=±—±-y_,j \sj \*sj

(All concentrations are expressed in moles per litre).

(28)

Temperature

Stumm and Lee (1961) observed that at constant pH and oxygen concentration, the rate of oxygenation increases tenfold for a 15°C increase in temperature. However, when Sung and Morgan (1980) normalised the experimental data with respect to changes in Kw and 02 solubility,

the rate constant varied only slightly with increasing temperature.

Organic matter

Iron can be complexed by humic and fulvic acids and similar organic substances present in water. Such complexation may render the iron resistant to oxidation even in the presence of dissolved oxygen (Oldham and Gloyna 1969; Jobin and Ghosh 1972; Knocke et al. 1992, 1994). Theis and Singer (1973,1974) showed that iron(II) complexation by humic matter increases with the increase of organic matter concentration and with the increase in pH. Furthermore, their study also showed that humic substances are capable of reducing iron(III), which depends on both the pH and the relative concentration of humic substances to iron(III).

Catalytic effect of some elements

Different ions present in water can alter the rate of homogeneous oxygenation of iron(II). Accelerating effects have been observed for Cu2+, Mn2+, Co2+ and H2P04" (Stumm and Lee 1961)

while inhibiting effects have been reported for SO42" and CI" (Sung and Morgan 1980). Barry et al. (1994), in their work on iron oxidation kinetics in aquatic ecosystems, also found that Ti02

accelerates the oxidation of iron(II) by forming complexes with it.

Silica in groundwater could interfere with the hydrolysis of oxidised iron and thus hinder filtration (Robinson 1975). Schenk and Weber (1968) reported that dissolved silica (H4Si04 or

Si(OH)4) affects the chemical behaviour of Fe2+ by catalyzing the rate of iron(II) oxidation. Dart

and Foley (1970) present some operational experiences that appear to be opposite to the conclusions drawn by Schenk and Weber (1968). Iron removal problems were experienced in waters with 30 or 40 mg/1 silica that often released very little of their iron content on either aeration or chlorination followed by filtration. The silica apparently reacts with Fe(OH)3 and

holds it in suspension (See Table 1.2-3. Stabilisation).

1.3.3 Hydrolysis of iron(III)

Iron(III) formed on oxidation of iron(II) subsequently undergoes hydrolysis resulting in the formation of hydrated iron oxide (Fe203.xH20). The aqueous chemistry of iron is rather complex

since this metal enters into several protolysis and oxidation-reduction reactions (Hem and Cropper 1962; Faust and Aly 1998; Stumm and Morgan 1996). Singley et al. (1967, 1969) demonstrated the existence of an entire family of iron hydrates having up to six molecules of water associated with one Fe3+ ion.

(29)

Lerk (1965) suggested that the oxidation and hydrolysis reactions of iron(II) occur as follows: 4Fe2+ + O2 + 2H20 -> 4Fe3+ + 40H " (1.9) 4Fe3+ + 40H " + 2(x+ 1)H20 -> 2(Fe203 xH20) + 8lf (1.10) or at high pH: 4Fe2+ + 80H'-> 4Fe(OH)2 (1.11) 4Fe(OH)2 +02+ (2x- 4)H20 -> 2(Fe203xH20) + 8tf (1.12)

The overall reaction can be written as

4Fe2+ + O2 + (2x + 4)H20 -^ 2(Fe203 xH20) + 8lf (1.13)

The hydrolysed species of iron ions will condense to form dimers through hydr and oxo-bridging. These are called "olation" and "oxolation" respectively. The dimers may undergo additional hydrolytic reactions that could provide additional hydroxo groups, which then could form more bridges. These processes lead to the formation of polynuclear hydroxy complexes and ultimately to the formation of precipitates.

Olation (hydroxo-bridging)

2[Fe(H20)5OH2+] -> [(H20)4Fe - (OH)2 - Fe(H20)4]4+ + 2H20 (1.14)

Dimer Oxolation (oxo-bridging)

2[Fe(H20)5OH2+] -* [(H20)5Fe - O - Fe(H20)5]4+ + H20 (1.15)

Dimer

1.3.4 Chemical oxidation

The atmospheric oxygen, which is introduced into water during aeration, is usually effective in the oxidation of iron(II). However, when the iron is organically complexed, aeration alone is not sufficient. Secondly, iron oxidation is very slow at pH <7.0. Alternative oxidants like potassium permanganate, chlorine or chlorine dioxide, ozone and hydrogen peroxide could be employed for iron oxidation when aeration is not satisfactory. Chemical oxidation is frequently applied when iron and manganese are to be removed simultaneously in a single filtration step as manganese oxygenation is very slow at pH <9.5.

The oxidation of iron(II) by different oxidants can be described by the following chemical reactions:

3Fe2++ KMn04 + 7H20 -> 3Fe(OH)3 + Mn02 + JC + SIT (1.16)

2Fe2+ + Ch + 6H20 -> 2Fe(OH)3 + 2Œ+ 6tf (1.17)

(30)

2Fe2+ + H202 + 4H20 -> 2Fe(OH)3 + 41? (L 1 9)

2Fe2+ + O3+ 5H20 -> 2Fe(OH)3 + 02+4F? O -20)

In each of the above cases, hydrogen ions are produced. Thus, based on these reactions alone, pH may decrease significantly in the absence of sufficient buffer capacity.

Very little has been reported in literature concerning the kinetics of oxidation of iron(II) with chemical oxidants (Willey and Jennings 1963; Abukhudair 1989; Knocke et al. 1991). As the chemical oxidation of iron is quite rapid at a pH of 7 or higher, the kinetic considerations have little influence on either facility design or operation (Benefield and Morgan 1990).

1.4 IRON REMOVAL MECHANISMS IN FILTERS

Different mechanisms (physical, chemical, and biological) may contribute to iron removal in filters but the dominant one depends on the physical and chemical characteristics of the water and process conditions (Lerk 1965; Rott 1985; Hatva 1988,1989; Mouchet 1992; S0gaard et al. 2000).

1.4.1 Oxidation-floc formation

Oxidation-floc formation (floe filtration) is the conventional approach for iron removal from groundwater. In this method soluble iron(II) present in anoxic groundwater is oxidised to insoluble iron(III) and after precipitation, iron hydroxide floes are removed in the filters. The removal process consists of the following steps (Rott 1973):

1. Oxidation of Fe2+ to Fe3+ by aeration or by a chemical oxidant

2. Hydrolysis of Fe3+ to iron hydroxides

3. Flocculation/agglomeration of the hydroxide particles. 4. Removal of floes in rapid sand filters.

This process is pH-dependent and dominant at pH values above 8.5. Under this condition, the oxidation is rapid and floes are formed prior to entering the filter bed.

Various problems have been encountered in the application of this mechanism. In some plants complete oxidation is not achieved, whereas in others filterable floes (precipitates) could not be formed (O'Connor 1971; Mouchet 1992). Dissolved iron(II) remaining and colloidal iron(III) formed can both pass the filter, consequently lowering the efficiency of iron removal. The rapid head loss development due to clogging of filters and rapid deterioration of filtrate quality are often responsible for short filter runs and frequent backwashing cycles of iron removal plants. Additionally, filter ripening after backwashing takes a rather long time and a large volume of sludge is produced which must be treated and/or disposed of.

(31)

The problem is severe when iron is organically complexed as aeration alone can not oxidise the complexed iron. To overcome the problems associated with oxidation of iron by aeration, strong oxidising agents such as chlorine, chlorine dioxide, potassium permanganate and ozone can be used. These chemicals also need sufficient detention time to allow the oxidation reaction to complete and, in addition, some of them can form unwanted by-products (Culp 1986). Provision of sufficient detention time also requires increased construction costs of the plant. Moreover, if the groundwater pH is low, chemicals are required to raise the pH and to enhance the oxidation, again associated with increased operation and maintenance costs.

1.4.2 Adsorption-oxidation

In the adsorption-oxidation (adsorptive filtration) mechanism, the iron(II) present in anoxic groundwater is removed by adsorption onto the surface of the filter media. Subsequently, in the presence of oxygen, the adsorbed iron(II) is oxidised forming a new surface for adsorption. In this way the process continues. The method therefore relies on the iron(II) adsorption capacity of the filter media. In conventional filters, the iron entering the filter bed in iron(II) form is removed through the adsorption-oxidation mechanism. Iron(II) can also adsorb on iron hydroxide floes commonly present in the filter. Adsorption-oxidation is also the dominant iron removal mechanism in dry filters and sub-surface iron removal (van Beek 1983; Rott 1985; Braster and Martinell 1988; Appelo et al. 1999). For the adsorption mechanism to dominate, pre-oxidation of iron(II) before filtration must be minimal. This can be achieved by reducing the oxidant concentration or the time available for the oxidation reaction. It should also be noted that adsorptive iron removal is only feasible for the removal of iron(II).

To achieve principally adsorptive iron removal, the filters can be operated in the following two modes:

(a) in intermittent regeneration mode, filters are operated under anoxic conditions. Oxidation of iron(II) is consequently suppressed by avoiding aeration. After the exhaustion of the iron(II) adsorption capacity of the filter media, the anoxic bed requires regeneration of the adsorption sites by oxidation of adsorbed iron(II). This can be achieved by backwashing the filter with oxygen-rich water or with a chemical oxidant e.g. KMn04;

(b) in continuous regeneration mode, filters are operated under aerobic conditions to allow continuous regeneration of the exhausted adsorption sites. A low concentration of oxygen and/or a short pre-oxidation time is required to avoid the formation of iron hydroxide floes. In this mode there are three possible options: a) dry filter, b) normal rapid filter operated at a high filtration rate and low depth of supernatant, and c) normal rapid filter with low oxygen concentration (1-2 mg/1) in the feed water. Some iron floes, however, will also be formed under these conditions and therefore backwashing will be required when maximum head loss is reached.

Many researchers studied the functioning of iron removal plants and observed that the iron oxide coating often plays an important role in the oxidation and removal of iron (Hauer 1950; Cox 1964; O'Connor 1971; Anderson et al. 1973). This was evident from the improved iron removal

(32)

after a coating of iron oxide had developed on the filter media. Some researchers have termed this as "Catalytic Iron Removal" because the iron removal rate increases with the formation of coating as the previously retained iron oxide acts as the "catalyst" for further iron removal. Cox (1964) found that filters may serve as contact beds following aeration where "catalytic action" of previously precipitated iron oxides will facilitate the oxidation of iron. From their study of iron removal in filters, Ghosh et al. (1967) concluded that a fraction of the ferrous iron might have been adsorbed onto the ferric hydrate precipitates. O'Connor (1971) reported that precipitates of hydrous oxides of iron(III) formed after oxidation and deposited on the filter sand serve as adsorption media for iron still in solution. The iron hydroxides have high adsorption capacities for iron(II), thus accounting for the improved removal when filters are ripened and deposition of the precipitates have taken place.

Previous researches at IHE (Adekoya 1995; Amoateng 1996; Ibrahim 1997) showed that water quality improves, filter ripening time reduces, and filter run time increases when iron(II) ions are removed via adsorption onto filter media, compared when iron(III) are removed as floes. Ibrahim (1997) reported that iron oxide coated sand has much higher efficiency for iron removal compared to new sand. Further, it was found that once the coating is developed, the filter media could give a higher efficiency even at lower pH. Sharma (1997) found that compared to new filter sand, iron oxide coated sand has very high capacity for iron(II) adsorption. This indicates the possibility of improving efficiency of iron removal in the filter by maximising the adsorption of iron(II) onto iron oxide coated media. Adsorption of iron(II) onto iron oxide coated media could be the primary iron removal method for treating anoxic groundwater and an attractive alternative to the conventional oxidation-floc formation method. With the adsorption-oxidation mechanism, the head loss is likely to be very low because the iron forms a coating on the filter media rather than a floe which blocks the filter pore. Thus, the filter runs could be longer and the backwash water requirement and volume of the sludge reduced. Under this mechanism, it is

ikely that filters could be run at higher filtration rates as head loss development is not a imitation. Hence, considerable savings are likely in the capital, and operation and maintenance (O&M) costs.

1.4.3 Biological iron removal

!ftl

C

H

a , 1

\

m

t

a t e d

,r

i d a t i 0 n 3nd rem0Val

°

f i r

°

n haS been re

P°

rted in

K * «»d ™&"

1 m uZZ g o ^ ". ai 1 9 8 5 ; C z e k a l l a et al m5> B a dJ° a n d douchet 1989; Hatva

c^vhks of mi ^

8me V1 1994)

-

Bi0l0giCal i r

°

n removal main

»y Spends on the

tSZL ZTTT

S

'

WhlC

,

h haVC the UDiqUe Pr0per

*

of c a u s i

i Nation and

SSbSLSS, r r f

PH

'

and red0X POtential

<

Eh

) « é t i o n s that are

S S f ^ T Tv-

na

^

ral

f

oundwater a

"

d

those required for conventional

(physical-* e ~ ^ ^

(33)

0 . 7 - 1

Stability gf ferrous Iron

Fig. 1.3 Field of activity of iron bacteria (Mouchet 1992)

The exothermic oxidation of iron(II) can be catalysed by some bacteria due to the oxidation-reduction enzymes which they excrete (flavins); trivalent iron rendered insoluble in hydroxide form is then stored in the mucilaginuos secretions (sheaths, stalks, capsules etc.) of these bacteria. The organisms responsible for this phenomenon are Gallionella, Leptothrix, Crenothrix,

Clonothrix, Siderocpasa, Sphaerotilus, Ferrobacillus and Sideromonas (Degremont 1991).

These iron-oxidising bacteria are widespread and are prevalent in groundwater, ponds, hypolimnion of lakes or impoundments, sedimentary deposits and soil. Two mechanisms of bacterial oxidation have been reported (Czekalla et al. 1985; Bourgine et al. 1994):

i) Intracellular oxidation by enzymatic action {Gallionella and Leptothrix ochracea), ii) Extracellular oxidation by the catalytic action of excreted polymers {Gallionella,

Leptothrix, Crenothrix, Clonothrix, Sphaerotilus, and Siderocapsa)

A pH of 6-8 is required for their activity. However, at a pH above 7.2, biological processes will compete with conventional (physical-chemical) processes. The optimum temperature typically ranges from 10 to 15°C for Gallionella ferruginea and 20 to 25°C for the Sphaerotilus-Leptothrix group (Mouchet 1992).

Mouchet (1992) reported the marked improvement in performance by converting conventional iron treatment plants to biological ones. The primary advantages associated with this process are

(34)

high filtration rates (10-70 m/h), high retention capacity (1-5 kg Fe/m2), elimination of chemical

reagents, flexibility of operation, and reduced capital and operating costs. However, high iron concentrations in the influent may cause breakthroughs, as the rate of absorption by the bacteria may not be high enough to match the supply rate.

The main disadvantage of this process is the long maturation time before full efficiency is achieved; perhaps 50-60 days for a new filter and 5 days after a 2-month shut down (Stevenson 1997). The other shortcomings of this mechanism include:

• anaerobic conditions may develop in the filter bed, thus converting back iron(III) to iron(II) resulting in an elevated iron concentration in the filtrate;

• increased sludge production and backwash water with filter ageing; • not suitable for all types of groundwater;

• need for two filtration stages to remove iron and manganese as the required redox potential conditions for iron and manganese oxidising bacteria are very different-. ineffective in the presence of ammonia (NlV) and inhibiting substances like H2S and Zn

(Tworte/a/. 1994; Stevenson 1997).

l l f v d,fated W hfei "n e a r n e U t r a l P H b a C t e r i a" a c t u a»y o x i d i s e ™ and grow

L t o l t y °r T " d e P°S i t i r°n " a n °X i d i S e d f 0 m- I n i r o n b a c t e r i a other than the

rmon^IT' ! Î T

ng fOT0US ir0n aS a S0UrCe 0f e n e r

gy

has n

<* been conclusively

EnTronme f TF? ™ ^ ^ * * * * d° h 0 W e V e r p r°C e S S i r o n -tracellularly

m e c h a Z f n § " ""? ^ l e a d t 0 d e V a t e d l e V e l s i n s i d e t h e «*• ™e cells have

^nSrlnTT,

unwanted ions

-

Therefore

'

oxidation of iron

™y

be

^ y

t o

^ ^ J S ^ ? " "T111116111 r a t h e r t h a n t 0 C r e a t e e n e r^ - T h e c e l 1 membrane and

ÏÏ^SZ8 t t """"H 6 m a n y SiteS for t h e a d S O r p t i°n °f i r 0 n<n) i o n s- Once adsorbed

he t h s 0ft n c h e d T ** °X l df ™i o n- ™s ^ads to the formation of the characteristic

^ Ï Î ^ H T ^

giCal i r

°

n removal U appears that bacteria can act as a

A£SÏÏT

Vla adsorption onto the cel1 membrane

°

r Via oxidation to

Ä « 2 ^ ^ T T Vann0t bC e S t a b H s h e d i f S u c h m e c h a n i s™ Fovide a

Mgnmcant removal capacity (Hughes and Poole 1989).

o a E l t e Ledit ÏuoZl T Ï T ^ ^ b y a d S O T p t i o n o n t o n e w °r ™ «-de

process Z ^ ^ ^ l V * f\ " ^ °f ^ ^ ^ ^ w M c h i n d i c a t e s t h a t t h e

complete « J ^ S ^ S ^ Ï """•*" T™* ^ ^ P Ü O t S ^ S n ^

fact that iron was reiXed at the ver! h ^ ^ ^ ^ the beginning of the test. The

this process is S Ä ™ W T ^ ?** *""** P r°C C S S S t r o n^ s u^ t s that

Plants the r m ^ ^ ^ ^ ^ f ^ T f ^ » ha* been reported that in some

the surface of the m ^ ^ ^ t ^ ^ T o f b i° **» «r "coatings" on

increased adsorption of ironfino^o \ ' , ^ " * l"4)' T h i s m ay b e * * to the

» « ^ o r b o J l ^ ^ ^ ^ ^ t h e a d d i t i- a l iron(II) oxidation by iron

i s p h y s i c o c h e m i c a l o r b ^ S